Global banks including Barclays (NYSE: BCS) and Deutsche Bank (NYSE: DB) are integrating Ant International’s FalconTST AI model to strengthen cash-flow predictions and foreign-exchange liquidity management.
In a development underscoring the growing role of specialised artificial intelligence in financial services, several international banks have begun incorporating a sophisticated forecasting tool developed by Singapore-based Ant International.
Institutions such as Barclays, Deutsche Bank, Citi and Standard Chartered are applying the Falcon Time-Series Transformer (FalconTST) model—specifically its upgraded 2.0 version—to refine predictions of cash flows and manage foreign-exchange (FX) liquidity more effectively amid the complexities of cross-border transactions.
The FalconTST system is a purpose-built time-series foundation model designed for sequential financial data rather than general language tasks.
It analyses patterns in transaction volumes, currency positions, account balances and related variables to generate forecasts on hourly, daily and weekly horizons.
Originally deployed within Ant International to oversee the company’s own cash-flow and currency exposure needs, the technology has now been embedded into the banks’ proprietary platforms.
Barclays, for instance, has linked it to its BARX NetFX system, while other partners have adapted it for their respective FX hedging and liquidity tools.
According to details shared by Ant International, the latest iteration achieves strong results on industry benchmarks, including a Mean Absolute Scaled Error (MASE) score of 0.666 that places it at the forefront of comparable time-series models.
Forecast accuracy is reported to exceed 93 percent in relevant applications.
This level of precision supports more informed decisions around liquidity preparation, currency hedging and capital allocation, potentially reducing associated costs.
Company representatives have noted that accurate projections can lower FX hedging and related expenses by more than 60 percent in certain scenarios by minimising unnecessary buffers and risk premiums.
The adoption reflects broader industry efforts to harness specialised AI for operational resilience.
Global FX markets handle enormous daily volumes, and volatility in currency rates can significantly affect businesses engaged in international trade, e-commerce and payments.
By integrating the model, the banks aim to provide enhanced forecasting capabilities both for their own operations and for clients, particularly those in high-volume cross-border environments.
Ant International has indicated that the technology is also being explored for adjacent sectors such as e-commerce demand planning, logistics and aviation operations management.
Executives from Ant International have emphasised the practical value of converting predictive insights into actionable strategies.
The model’s focus on financial scenarios gives it an edge over more generalised large language models that have not yet delivered comparable breakthroughs in this domain.
Kelvin Li, general manager of platform technology and senior vice president at Ant International, highlighted how improved accuracy extends benefits to banking partners and customers in dynamic industries.
Jiang-Ming Yang, chief innovation officer, pointed to the importance of translating forecasts into concrete choices on liquidity, exposure and capital use.
This collaboration builds on earlier individual partnerships between Ant International and various banks, which began with pilots and initial integrations focused on treasury efficiency.
The coordinated uptake of the refined 2.0 version signals accelerating momentum toward AI-driven risk management in global finance.
As financial institutions continue to prioritise technology that addresses rapid shifts in liquidity needs and currency movements, tools like FalconTST illustrate how targeted predictive systems can support more confident handling of international payment flows and treasury operations.
While specific implementation details and long-term outcomes will vary by institution, the move highlights a clear trend: specialised AI is becoming an integral component of modern banking infrastructure for cash-flow and FX challenges.